Mesencephalic astrocyte-derived neurotrophic factor reprograms macrophages to ameliorate acetaminophen-induced acute liver injury via p38 MAPK pathway

Acetaminophen (APAP)-induced liver injury (AILI) is the most frequent cause of acute liver failure; but the underlying mechanisms still remain obscure. Macrophages and endoplasmic reticulum (ER) stress play an important role in the pathogenesis of AILI. Mesencephalic astrocyte-derived neurotrophic factor (MANF) is a newly identified 18-kDa soluble protein, whose expression and secretion are stimulated by ER stress. To investigate the role of myeloid cell MANF in the pathogenesis of AILI, we assayed serum and liver samples from AILI model mice and patients with drug-induced liver injury (DILI). We demonstrated that the levels of MANF were elevated in patients with DILI and in mice with AILI. Moreover, myeloid-specific MANF knockout mice were generated and used. It was observed that a delayed liver recovery from myeloid-specific MANF gene knockout mice following APAP overdose compared to that from wild-type mice. MANF deficiency in myeloid cells resulted in increased infiltrating monocyte-derived macrophages (MoMFs) but reduced restorative Ly6Clow macrophages after APAP treatment. MANF supplementation increased restorative Ly6Clow macrophages and subsequently alleviated liver injury. Moreover, MANF could enhance IL-10 expression and phagocytosis in macrophages via p38 MAPK pathway. Altogether, MANF seems to be a critical immune modulator in promoting liver repair via reducing and reprogramming MoMFs. MANF perhaps promoted the phenotype conversion of pro-inflammatory MoMFs to pro-restorative Ly6Clow MoMFs via p38 MAPK pathway, particularly through enhancing IL-10 and phagocytosis.


INTRODUCTION
Drug-induced liver injury (DILI) is one of the major causes of liver diseases around the world. In Western society, the main cause of acute liver injury is acetaminophen (APAP) overdose [1]. APAP is metabolized by hepatic cytochrome P450 enzymes (mainly CYP2E1) into a reactive metabolite N-acetyl-p-benzoquinoneimine (NAPQI) which is detoxified by glutathione (GSH) [2]. Excessive formation of NAPQI causes depletion of hepatic glutathione in APAP overdose and subsequently leads to oxidative stress and hepatocyte damage [3]. Currently, N-acetylcysteine is the only clinically approved antidote for APAP overdose patients. However, the efficacy of N-acetylcysteine dramatically decreases 8 h after APAP ingestion [4,5]. Therefore, the development of new approaches to treat APAP-induced liver failure, particularly at the late phase of injury is required.
Initial NAPQI-induced direct hepatocyte damage leads to the release of damage-associated molecular patterns (DAMP), which activates cytokines and chemokines formation in resident Kupffer cells for recruitment of neutrophils and circulating monocytes [6,7]. CCL2 (MCP-1) provokes recruitment of circulating CCR2 + Ly-6C hi monocytes into injured liver [8]. These monocyte-derived macrophages (MoMFs) represent the main liver macrophage population after APAP injury, and play contradictory roles in APAP-induced liver injury (AILI) [9]. It has been demonstrated that Ly-6C hi macrophages have a pro-inflammatory phenotype reflected by the expression of inflammatory cytokines. Hence, antagonists against CCL2 or CCR2 ameliorates the early but not the late phase of AILI via reducing the recruitment of monocytes [10,11]. Interestingly, other studies report that these Ly-6C hi monocytes differentiate into Ly-6C low restorative macrophages at the site of injury. Ly-6C low macrophages express anti-inflammatory cytokines, regenerative growth factors and matrix degrading metalloproteinase to promote resolution from injury [12][13][14][15]. Signals derived from the hepatic microenvironment (e.g. macrophage colony-stimulating factor 1 (CSF1), IL-6, phagocytosis of cellular debris) may promote the phenotypic and functional transition of macrophages [16][17][18].
Mesencephalic astrocyte-derived neurotrophic factor (MANF), a new evolutionary conserved neurotrophic factor, is widely expressed in neuronal and non-neuronal tissues [19,20]. MANF expression and secretion are regulated by endoplasmic reticulum (ER) stress and protects against various ER stress-induced damage [21,22]. Recent studies demonstrate that MANF has an autocrine immunomodulatory function in macrophages, which is essential for its neuroprotection and tissue repair in retina and liver [23]. However, the underlying mechanism of MANF in signaling remains elusive. In this study, we identified MANF to be a protective mediator that promoted resolution of AILI by regulating monocyte infiltration and triggering macrophages conversion into a reparative phenotype. Mechanistically, we demonstrated that MANF promoted the production of IL-10 and phagocytosis in macrophages though p38 kinase. Treatment of mice with recombinant human MANF (rhMANF) was sufficient to facilitate liver repair, which suggested that MANF may serve as a promising therapeutic strategy for APAP overdose and other liver injuries.

MATERIALS AND METHODS Human samples and ethics statement
Liver tissues and blood samples of patients with drug-induced liver injury were collected from the First Affiliated Hospital of University and Science and Technology in this study (n = 24). In addition, 24 healthy blood samples were included as a control group. Normal liver samples were obtained from surgical resections of uninvolved liver tissues of angioma patients (n = 20). Informed consent was obtained from all subjects. This study was approved by the Ethics Committee of Anhui Medical University and the First Affiliated Hospital of University and Science and Technology.

Animals
Manf flox/flox mice with C57BL/6 J background were kindly provided by Professor Jia Luo of Kentucky University, USA. Manf Hep-/mice and Manf Mye-/mice were generated by a few steps including crossing Manf flox/flox mice with Albumin Cre mice and Lysosome Cre mice, respectively. All mice were maintained in a specific pathogen-free facility. All animal experiments were approved by the Institutional Animal Care and Use Committee at Anhui University.

Animal treatment
To induce acute liver injury, fresh solutions of APAP (Sangong, Shanghai, China) were prepared immediately before use by dissolving the compound in warmed PBS. After overnight fasting, 8-10 week old male mice were injected with APAP 300 (mg/kg) intraperitoneally. For the treatment, some of the mice were intravenously injected with recombinant human MANF (1.5 mg/kg, SinoBiological, Beijing, China) or recombinant mouse IL-10 (1 μg/mouse, Biolegend, San Diego, CA, USA) 0.5 h after APAP administration.

Serum ALT assay
Serum ALT levels were measured using an ALT assay kit (Nanjing Jiancheng Bioengineering Institute) according to the manufacturer's instructions.

Cell isolation
Hepatic leukocytes were isolated as described previously [24]. Briefly, liver tissues were minced and sieved through a 70 μm filter. Hepatic leukocytes were purified by centrifugation on a 40% Percoll gradient (GE Healthcare), and red blood cells were lysed by RBC lysis buffer. Collected leukocytes were further enriched using PE-anti-F4/80 (BD Pharmingen, Cat# 565410) and anti-PE microbeads (Miltenyi Biotec) to isolate macrophages. Hepatocytes were isolated as described previously [25]. Briefly, the liver was perfused with a solution of EGTA and digested with a 0.075% collagenase solution. The viable hepatocytes were separated by 40% Percoll solution with centrifugation at 420 × g for 10 min at 4°C.

Histology assay
Liver tissues were fixed in 10% formalin and paraffin-embedded following standard procedure. Embedded liver tissues were sectioned, and stained with hematoxylin and eosin (H&E) or subjected to immunohistochemical staining using anti-MANF antibody (Abcam, Cambridge, MA, USA, ab67271), anti-Cyp2E1 antibody (Abcam, Cambridge, MA, USA, ab28146) according to the manufacturers' protocols. The necrosis areas were outlined according to the cellular structure and color. The percentages of necrosis areas were averaged from 5 to 7 random view fields of each section (×200 magnification) using ImageJ software.

Phagocytosis assays
For detecting phagocytosis of macrophages, mouse primary hepatocytes or human hepatocytes (LO2 cells) were treated with 20 μM dexamethasone for 24 h and then labeled with 1 μM CFSE for 15 min at 37°C. Mouse peritoneal macrophages, liver MoMFs (isolated from the same mouse as primary hepatocytes), or human monocytes (THP-1 cells) were challenged with CFSE-labeled apoptotic hepatocytes for 1 h at a 1:5 ratio at 37°C or X. Hou et al.
4°C. The cells were washed and stained with cell surface markers. Stained cells were analyzed by flow cytometry.

Statistical analysis
Data are expressed as the means ± SD and were analyzed using GraphPad Prism software (v. 8.0a; GraphPad Software, La Jolla, CA, USA). The variance was similar between the groups that were being statistically compared. Comparisons between two groups were performed using an unpaired Student's t-test. Comparisons among multiple groups were performed using two-way ANOVA followed by Tukey's post hoc test. P < 0.05 were considered significant. Sample sizes of all experiments were predetermined from preliminary experiments or from reports in the literature. No sample was excluded from the analyses. All mice used in this study were chosen randomly. Investigators were not blinded to the group allocation during the experiment and outcome assessment.

Increased MANF expression in DILI patients and an AILI mouse model
To investigate the role of MANF in DILI, we first examined MANF expression in DILI patients. As shown in Fig. 1A, B, serum levels and hepatic levels of MANF protein were markedly increased in DILI patients compared to healthy controls. A similar increase in MANF levels in the serum and liver tissues was observed in the AILI mouse model (Fig. 1C, D). Moreover, we performed an Immunofluorescence analysis to co-stain hepatic macrophages and hepatocytes with MANF. As illustrated in Fig. 1E, the expression of MANF in F4/80 + macrophages was increased significantly after APAP challenge, whereas the expression of MANF in hepatocytes was increased mildly (Supplementary Fig.  S1). Consistent with the protein analysis results, the increase of Manf mRNA levels in macrophages (about 3-fold) was higher than that in hepatocytes (about 1.3-fold) (Fig. 1F).

MANF in myeloid cells is critical for resolution of AILI
To further dissect the role of MANF in hepatocytes and myeloid cells, we generated hepatocyte-specific Manf knockout mice (Manf Hep-/-), and myeloid-specific Manf knockout (Manf Mye-/-) mice, and treated these mice with APAP. Compared to WT mice, Manf Mye-/mice showed higher levels of serum alanine aminotransferase (ALT), larger areas of liver necrosis, and lower numbers of Ki67 + hepatocytes at the late phase post-APAP injection, whereas Manf Hep-/mice had comparable liver injury and regeneration to WT mice ( Fig. 2A-D). In addition, deletion of MANF in hepatocytes did not significantly affect APAP-induced hepatocyte death in vitro ( Supplementary Fig. S2). These results suggested that myeloid MANF played an important role in liver injury resolution after APAP treatment.
To study whether the delayed resolution of AILI in Manf Mye-/mice was attributed to alteration of APAP metabolism, we examined CYP2E1 and GSH expression in the liver. Our results showed that liver CYP2E1 protein expression and hepatic GSH levels were comparable between WT and Manf Mye-/mice after overnight fasting or nonfasting ( Supplementary Fig. S3A-C).
Manf Mye-/mice have increased MoMFs but reduced restorative Ly6C low MoMFs after APAP treatment We then evaluated the populations of macrophages and neutrophils in the liver of Manf Mye-/mice after APAP treatment by performing flow cytometric assays. The gating strategies for liver neutrophils and MoMFs were shown in Supplementary Fig. S4A. The percentages of neutrophils (CD45 + CD11b + Ly6G + ) ( Supplementary  Fig. S4B, C) and MoMFs (CD45 + CD11b high Ly6G -F4/80 intermediate ) (Fig. 3A) were increased in the liver of Manf Mye-/mice compared with WT mice at the late phase of APAP treatment. Next, we found that hepatic Ccl2 and Ccl3 levels were higher in Manf Mye-/mice after APAP treatment than those in WT mice ( Supplementary Fig. S5A). Furthermore, Manf Mye-/mice showed increased monocytes in peripheral blood after APAP treatment compared to WT controls ( Supplementary Fig. S5B, C). We also observed that monocytes from Manf Mye-/mice had a higher expression of CCR2 ( Supplementary  Fig. S5D, E). In addition, MANF deficiency did not influence the proliferation and survival of macrophages ( Supplementary Fig.  S6A-D). These results suggested that MANF limited monocyte influx by inhibiting the expression of Ccl2 and Ccl3 in the liver and the expression of CCR2 on monocytes.
Moreover, Manf Mye-/mice had a lower percentage of Ly6C low restorative MoMFs and a higher percentage of Ly6C hi MoMFs between 24 and 72 h post-APAP application (Fig. 3B). To further confirm the protective role of Ly6C low restorative MoMFs in AILI, we adoptively transferred Ly6C low MoMFs from APAP-treated WT mice into Manf Mye-/mice at the same time with APAP administration. The Manf Mye-/mice that received Ly6C low MoMFs showed decreases in ALT levels and hepatic necrosis (Fig. 3C, D).
rhMANF enhances restorative macrophage percentage in vivo and facilitates liver recovery after acute liver injury The beneficial function of MANF in liver injury resolution prompted us to test whether MANF delivery is sufficient to elicit protective effects; thus, we treated WT mice and Manf Mye-/mice with rhMANF at the same time as APAP administration. As shown in Fig. 4A, rhMANF treatment caused remarkable reductions in MoMFs and pro-inflammatory Ly-6C hi MoMFs and a significant increase in restorative Ly-6C low MoMFs during liver injury resolution (48 h after APAP administration). Critically, rhMANF treatment greatly decreased ALT levels ( Supplementary Fig. S7A and Fig. 4B) and hepatic necrosis ( Supplementary Fig. S7B, C and Fig. 4C, D) in both WT mice and Manf Mye-/mice 24 h and 48 h after APAP administration. These data indicated that rhMANF induced macrophages to switch to a phenotype promoting liver injury resolution in vivo.
The effect of myeloid cell-derived MANF on AILI depends on IL-10 Given that Ly6C hi MoMFs secret inflammatory cytokines [10], whereas Ly6C low MoMFs secrete anti-inflammatory cytokines [7], we examined cytokine expression in the liver of APAP-treated mice. RT-qPCR analysis showed significantly increased hepatic Tnfa and Il1b levels and decreased Il10 levels in Manf Mye-/mice compared to WT mice after APAP treatment (Fig. 5A). Moreover, Il10 levels in hepatic macrophages of APAP-treated Manf Mye-/mice were much lower than those in APAP-treated WT mice, and rhMANF dramatically up-regulated Il10 levels in hepatic macrophages of Manf Mye-/mice (Fig. 5B). In addition, rhMANF increased Il10 levels in the liver of both WT mice and Manf Mye-/mice, whereas had no effect on hepatic Tnfa and Il1b levels ( Supplementary Fig. S8A).
It has been reported that macrophages secrete the antiinflammatory molecule IL-10 that actively promote tissue repair [18,26]. To explore whether delayed liver injury recovery is due to decreased IL-10 levels, we treated Manf Mye-/mice with IL-10 after APAP injection. As expected, treatment of mice with IL-10 significantly reduced liver injury at 24 h after APAP injection, which was demonstrated by reduced ALT levels (Fig. 5C) and less hepatocyte damage (Fig. 5D, E).

MANF induces IL-10 expression in macrophages via p38 MAPK
To understand the mechanism by which MANF regulates IL-10 expression in macrophages, we screened several potential signaling pathways that are involved in IL-10 production in myeloid cells [27]. Our data revealed that rhMANF activated AKT and the MAPK family, including JNK, p38, and ERK1/2 in peritoneal macrophages of WT mice (Fig. 6A). We then treated peritoneal macrophages with rhMANF in combination with an AKT inhibitor and various MAPK inhibitors. Interestingly, only p38 inhibitor reduced IL-10 expression (Fig. 6B). Furthermore, the phosphorylation of p38 in liver macrophages of Manf Mye-/mice was decreased compared with WT mice after APAP treatment (Fig. 6C). In vitro D Representative IHC images of hepatic MANF expression in mice 24 h after APAP treatment and in control mice. E Immunofluorescence staining for F4/80 (green), MANF (red) and DAPI for nuclei (blue) in WT mice liver tissue. Orange arrows point to F4/80 + macrophages with high expression of MANF. F Relative mRNA expression of MANF in macrophages and hepatocytes at 0 h (control) and 12 h after APAP treatment was determined by qPCR. Data are presented as mean ± SD and representative of at least 3 independent experiments. n = 6-8 per group. *P < 0 .05, **P < 0.01. Fig. 2 Manf Mye-/mice display delayed liver recovery after APAP treatment. WT mice, Manf Mye-/mice, and Manf Hep-/mice were intraperitoneally injected with APAP. A Serum ALT levels were measured after APAP treatment. B Representative H&E-staining and IHC staining for Ki67 in liver sections after APAP treatment. Necrotic areas were outlined using red lines. The percentages of necrosis area C and Ki67 D positive hepatocytes were quantified from 5 to 7 random view fields of each section (×200 magnification) using ImageJ software. n = 5-7/ group, 3 independent experiments. Data are presented as mean ± SD. *P < 0.05, **P < 0.01. experiments showed that when treated with LPS or rhMANF the phosphorylation of p38 was decreased in peritoneal macrophages of Manf Mye-/mice as compared with WT mice (Fig. 6D). These results suggested that MANF induced IL-10 expression by activating the p38 signaling pathway. To our surprise, rhMANF also induced Tnfa, Il1b, and Il6 mRNA expression in naive mouse peritoneal macrophages and human macrophages, although this effect did not selectively depend on p38 MAPK (Supplementary Fig. S8B and Supplementary Fig. S9A). Considering that liver macrophages may be stimulated by DAMP-derived from APAP damaged hepatocytes in vivo, we further examined whether rhMANF could downregulate macrophage-mediated inflammatory response. We found that rhMANF pretreatment inhibited LPSinduced Tnfa, Il1b, and Il6 expression, but not Il10 ( Supplementary  Fig. S8C and Supplementary Fig. S9B).
MANF promotes macrophage phagocytosis via the p38 signaling pathway It has been reported that phagocytosis drives the conversion of inflammatory macrophages to restorative macrophages [16,18]. We analyzed the role of MANF on macrophage phagocytosis. A number of phagocytosis-related genes were examined in hepatic macrophages derived from APAP-treated WT mice and APAP/ rhMANF-treated WT mice. As shown in Fig. 7A, rhMANF treatment upregulated the expression of Gpnmb, Cd5l, Macro, Mertk, Trem2, and CD81. The apoptotic hepatocyte uptake assay further demonstrated that rhMANF treatment greatly increased MoMF phagocytosis in vivo (Fig. 7B). Given that p38 has been reported to play a pivotal role in the ability of macrophages to phagocytose bacteria or particles [28,29], we then examined whether the p38 signaling pathway is involved in the MANF-induced increase in macrophage phagocytosis. Our results showed that the p38 MAPK inhibitor decreased rhMANF-induced phagocytosis-related gene expression and phagocytosis of apoptotic hepatocytes in mouse peritoneal macrophages (Fig. 7C, D). It is worth noting that recombinant IL-10 restored the decreased phagocytic ability of macrophages caused by p38 inhibitor (Fig. 7C, D).
MANF promotes IL-10 expression and phagocytosis in human monocytes through the p38 signaling pathway Consistent with the above findings, rhMANF treatment also activated AKT, JNK, p38, and ERK1/2 MAPK in human THP-1 monocytes (Fig. 8A). All of the kinase inhibitors including p38 inhibitor, significantly reduced IL-10 expression levels in THP-1 Fig. 4 The rhMANF enhances the restorative macrophage phenotype in vivo and accelerates liver injury resolution. WT mice and Manf Mye-/mice were intravenously injected with rhMANF (1.5 mg/kg) at the same time as APAP administration. Mice were sacrificed at 48 h after APAP administration. A The percentages and numbers of hepatic MoMFs, Ly-6C hi expressing MoMFs, and Ly-6C low expressing MoMFs were quantified by flow cytometric analysis. B Serum ALT activities were determined. C Representative H&E-stained liver sections. n = 5-7 sections/mouse. D Necrotic areas in the liver were quantified by ImageJ software. Data are presented as mean ± SD, n = 5-7 mice per group, 3 independent experiments. cells (Fig. 8B). In addition, p38 inhibitor decreased rhMANFinduced phagocytosis-related gene expression and the phagocytosis of apoptotic hepatocytes in THP-1 cells (Fig. 8C, D). Moreover, IL-10 restored the decreased phagocytic ability of THP-1 cells caused by p38 inhibitor (Fig. 8C, D).

DISCUSSION
Recently, MANF has been described to have a function in regulating immune homeostasis in aging and protecting against liver damage [22]. However, the role of MANF in acute liver injury is still largely unclear. In this study, we revealed a function of MANF in mediating tissue repair after AILI by limiting MoMF infiltration and triggered inflammatory Ly6C hi monocytes transition to pro-resolution Ly6C low MoMFs. Mechanistically, the interaction of MANF with p38 resulted in increased IL-10 expression and enhanced phagocytosis in macrophages. The fact that rhMANF treatment was sufficient to regulate MoMF infiltration, macrophage programming, and liver repair, suggesting the therapeutic potential of rhMANF in treating APAP overdose-induced liver injury and other liver diseases.
We demonstrated that MANF was highly elevated in DILI patients and in an AILI mouse model. Both hepatic macrophages and stressed hepatocytes expressed MANF in response to liver injury, however, the increase of MANF expression in macrophages was higher than that in hepatocytes. Furthermore, Manf Mye-/mice showed increased inflammation and liver injury at the late phase after APAP injection, whereas no increase in inflammation and liver injury was observed in Manf Hep-/mice compared to WT mice, suggesting that the autocrine immunomodulatory function of MANF in myeloid cells was essential for the resolution of APAP-induced liver injury. Notably, rhMANF treatment was sufficient to promote liver repair in both WT mice and Manf Mye-/mice. A recent study showed that MANF deficiency in Cx3cr1expressing monocytes or macrophages resulted in liver damage and inflammation, however, hepatocyte-specific ablation of MANF did not result in increased immune activation and liver damage [22], which supported our findings.
Our data showed that MANF could extensively inhibit monocyte infiltration during the resolution phase. Manf Mye-/mice had increased levels of CCL2 and CCL3 in the liver. MANF-deficient CD115 + monocytes expressed higher surface levels of CCR2, which augmented cell mobilization and influx, possibly explaining the increased monocyte infiltration during the late phase compared with WT mice. During AILI, bone marrow-derived monocytes are recruited in the damaged liver tissue [30]. Generally, the main purpose of these monocytes is to remove necrotic cell debris and promote liver repair, but dysregulated inflammation serves as a detrimental factor in many liver diseases by exacerbating liver damage [31]. Blocking recruitment of infiltrating Ly-6C hi monocytes during the fibrosis regression phase accelerates resolution of toxic and metabolic liver fibrosis [32]. In this study, rhMANF treatment inhibited the extinctive monocyte Fig. 6 MANF induces IL-10 expression in macrophages via p38 signaling pathway. A Primary peritoneal macrophages of naive WT mice were treated with 2 μg/mL of rhMANF. Activation of various signaling pathways was measured by western blotting. B Primary peritoneal macrophages of naive WT mice were pretreated with inhibitors for 1 h, followed by rhMANF treatment for 5 h. Il10 mRNA expression was measured by RT-qPCR. C Phosphorylated and total p38 levels in primary liver macrophages from WT and Manf Mye-/mice treated with APAP for 2 h. D Primary peritoneal macrophages derived from naive WT mice and Manf Mye-/mice were treated with 100 ng/mL LPS or 2 μg/mL of rhMANF for 0.5 h. Phosphorylated and total p38 levels were analyzed by Western blotting. Data are presented as mean ± SD. *P < 0.05, **P < 0.01. n = 3-6 mice per group, 3 independent experiments.
infiltration in the liver, which might partially contribute to the accelerated liver injury regression induced by rhMANF. Future studies will need to determine how MANF influences CCR2/CCL2 expression. It will also be interesting to discern whether neutrophils which are also deficient in MANF, contribute to the heightened influx of monocytes in Manf Mye-/mice.
On the other hand, MANF orchestrated liver repair by triggering hepatic macrophage transition into a reparative phenotype. First, MANF promoted MoMFs to produce IL-10, which is an antiinflammatory cytokine and was expressed much higher in Ly6C low macrophages than that in Ly6C hi macrophages [33]. Deficiency of the IL-10 gene in mice resulted in higher liver injury after APAP Fig. 7 MANF enhances macrophage phagocytosis via the p38 MAPK activation. A, B Liver macrophages were isolated from naive (control), PBS/APAP-or rhMANF /APAP-treated WT mice 48 h after APAP injection. The phagocytosis-related gene levels in liver macrophages were measured. B Hepatocytes isolated from WT mice were treated with 20 μM dexamethasone for 24 h and then labeled with 1 μM CFSE for 15 min at 37°C. Freshly isolated liver macrophages were incubated with CFSE-labeled hepatocytes at either 37°C or 4°C (control) for 30 min and analyzed by flow cytometry. Uptake of hepatocytes by MoMFs was shown as Δ percentage (%) of CFSE positive MoMFs at 37-4°C. C, D Peritoneal macrophages isolated from naive WT mice were pretreated with p38 inhibitor for 1 h, followed by rhMANF (2 μg/mL) with or without IL-10 (20 ng/mL) treatment for 5 h. C Phagocytosis-related gene expression was measured by RT-qPCR. *P < 0.05 compared with control; #<0.05 compared with rhMANF-treated group, & <0.05 compared with p38 inhibitor + rhMANF-treated group. D Uptake of hepatocytes by peritoneal macrophages were examined by flow cytometry and shown as Δ mean fluorescence intensity (MFI) at 37-4°C. Data are presented as mean ± SD. n = 4-6 mice per group, 3 independent experiments. *P < 0.05, **P < 0.01, ***P < 0.001. Fig. 8 MANF induces IL-10 expression and phagocytosis in human monocytes via the p38 signaling pathway. A THP-1 cells were cultured with rhMANF (2 μg/mL) for the indicated times. Activation of various signaling pathways was measured by western blotting. B THP-1 cells were pretreated with various inhibitors for 1 h, followed by rhMANF (2 μg/mL) treatment for 5 h. Il10 mRNA expression in THP-1 cells was measured by RT-qPCR. C, D THP-1 cells were cultured with p38 inhibitor for 1 h, followed by rhMANF (2 μg/mL) with or without IL-10 (20 ng/ mL) treatment for 5 h. C Phagocytosis-related gene expression was measured by RT-qPCR. *P < 0.05 compared with control; #<0.05 compared with rhMANF-treated group, & <0.05 compared with p38 inhibitor + rhMANF-treated group. D Uptake of CFSE-labeled LO 2 cells by THP-1 cells were examined by flow cytometry and shown as Δ mean fluorescence intensity at 37-4°C. Data are presented as mean ± SD. n = 3/group, 3 independent experiments. *P < 0.05, **P < 0.01. injection by enhancing the levels of NO and proinflammatory cytokines [34]. In our study, Manf Mye-/mice showed decreased levels of Il10 in the liver especially in hepatic macrophages. Treatment of Manf Mye-/mice with recombinant IL-10 protected against AILI. Second, MANF promoted macrophage phagocytosis. rhMANF treatment upregulated the expression of phagocytosisrelated genes in macrophages thus enhancing macrophage phagocytosis. It has been reported that induction of phagocytic behavior enhances the restorative phenotype of macrophages [16,18]. Phagocytosis of necrotic hepatocytes by macrophages is important for liver repair and regeneration [35].
Mechanically, our results demonstrated that MANF promoted macrophage transition into a reparative phenotype, possibly by interacting with p38 MAPK. The p38 MAPK pathway is well known for its role in transducing stress signals from environments such as pathogens, heat shock, growth factors, and cytokines [36]. In this study, MANF enhanced p38 phosphorylation in macrophages. Previous studies have demonstrated that p38 activation contributed to IL-10 production in TLR-stimulated macrophages [37,38], and enhanced the ability of macrophages to phagocytose bacteria [29,39]. In addition, p38/MAPKAP kinase 2 (MK2) has been shown to be involved in the repolarization of M2 tumor-associated macrophages [40,41]. A recent study showed that MANF protected mice against high-fat diet-induced obesity by promoting adipose browning via p38 MAPK pathway [42]. Here, our data showed that MANF promoted IL-10 expression and phagocytosis in macrophages via p38 MAPK. Notably, our in vitro experiments showed that IL-10 restored the decreased phagocytic ability of macrophages caused by p38 inhibitor, which suggested that MANF-induced IL-10 could further enhance the ability of macrophages to phagocytose necrotic hepatocytes.
More recently, a series of studies suggested that inflammatory Ly6C hi monocytes transition into restorative Ly6C low macrophages at the site of injury [12,14,16,30,43]. Identifying the molecular mechanisms underlying the phenotypic switch might serve as a starting point for new translational approaches. In this study, we provided evidence that myeloid-derived MANF was an important endogenous protective factor that promoting resolution of AILI. The rhMANF has therapeutic potential at the late phase of AILI. The protective role of MANF was mediated by inhibiting the extensive accumulation of monocytes in the liver and promoting macrophage transition into a reparative phenotype. The MANF and p38 were critical for shaping the phenotype and function of MoMFs.

DATA AVAILABILITY
The data that support the findings of this study are available from the corresponding author upon reasonable request.